Hi all,

My understanding of harmonics in general is quite sketchy and as a result my query might appear a bit lulu! I tried the net but could not get suitable info on the same.

This is what i'm looking at;
Assume a battery hooked up to an inverter (1 phase or 3 phase doesn't matter). The output of the inverter is connected to a typical AC load. I do not know the relation between nature of AC load (resistive, inductive or capacitive) and hence the resultant harmonics injected into the DC line due to the same, hence I would be unable to elaborate on the same.

What I want to understand is; whether rapid switching of the inverter inject harmonics into the pure DC line? If so, what is their nature and waveform like? I would be grateful if someone could provide me some articles or an on-line link on the same.

Thanks & Kind regards,
Shahvir

 

Yes, inverters create harmonics. Digital switching is used to approximate a sinusoid like the power company's 60 Hz. But it is only an approximation. Looking at a scope you'd see an irregular shaped sinewave, looking at a spectrum analyzer you'd see a whole bunch of harmonics with those closest to 60 Hz being larger than those farther away (for the most part).

 

Yes, inverters create harmonics. Digital switching is used to approximate a sinusoid like the power company's 60 Hz. But it is only an approximation. Looking at a scope you'd see an irregular shaped sinewave, looking at a spectrum analyzer you'd see a whole bunch of harmonics with those closest to 60 Hz being larger than those farther away (for the most part).

Thanx for your reply. But in order to understand harmonics injected into the DC line due to same, I'll need to look at some waveforms depicting the same for clarity.

Best regards,
Shahvir

 

Thanx for your reply. But in order to understand harmonics injected into the DC line due to same, I'll need to look at some waveforms depicting the same for clarity.

Best regards,
Shahvir

To do this, you need some type of model of your DC source. Perhaps the simplest model would be a ideal DC voltage source in series with a source resistance. Then put a filter capacitor in parallel. You could also include line inductance depending on the frequency and length of cables.

So basically, I think the harmonics injected on the DC side are due to the current changes with a dynamically changing load. As the current changes, the DC voltage is modulated as the current in the source resistor and source inductance changes. Of course, the filter capacitor helps.

Personally, I would make a simple model in Simulink or Spice etc, and evaluate this numerically. Perhaps there is a simple way to do this analytically with equations, but I'd have to think about it. Maybe someone else knows?

The following book is an excellent reference about PWM inverters. It goes extensively into the math of harmonics. However, I can't remember if it looks at harmonics on the DC side. I have the book at work, so I will check on Monday.

http://www.amazon.com/Pulse-Width-M...=sr_1_1?ie=UTF8&s=books&qid=1242484726&sr=8-1

 

Thanx, I'll keep posted.

 

Here's what it might look like on a spectrum analyzer. The center line is the fundamental while those to either side are unwanted harmonics. If only the fundamental existed you'd have a pure sinewave.

I was told by someone that you don't want to filter a switching supply for the fundamental because you're throwing away power. There's alot of power in those sidebands.

 

Here's what it might look like on a spectrum analyzer. The center line is the fundamental while those to either side are unwanted harmonics. If only the fundamental existed you'd have a pure sinewave.

Did you forget an attachment?

 

Hi all,

My understanding of harmonics in general is quite sketchy and as a result my query might appear a bit lulu! I tried the net but could not get suitable info on the same.

This is what i'm looking at;
Assume a battery hooked up to an inverter (1 phase or 3 phase doesn't matter). The output of the inverter is connected to a typical AC load. I do not know the relation between nature of AC load (resistive, inductive or capacitive) and hence the resultant harmonics injected into the DC line due to the same, hence I would be unable to elaborate on the same.

What I want to understand is; whether rapid switching of the inverter inject harmonics into the pure DC line? If so, what is their nature and waveform like? I would be grateful if someone could provide me some articles or an on-line link on the same.

Thanks & Kind regards,
Shahvir

Tell us more about this inverter. Is it's output a so-called "modified sine wave", which is really a square wave with some dead time? Or, is it outputting a fairly good sine wave, using PWM modulation? Or, is it the "stepped approximation" type (such as the Xantrex SW4024)?

Are you only interested in the high frequency harmonics, due to rapid switching edges, with frequencies up in the radio frequency range? Or, are you also interested in the harmonics of the line frequency (perhaps up to the 51st harmonic) caused by the inductance of the battery cable?

 

Here's what it might look like on a spectrum analyzer. The center line is the fundamental while those to either side are unwanted harmonics. If only the fundamental existed you'd have a pure sinewave.

I was told by someone that you don't want to filter a switching supply for the fundamental because you're throwing away power. There's alot of power in those sidebands.

Paul, if I understand you, you're talking about harmonics on the inverter output. I believe our OP is asking about harmonics (noise, actually - DC doesn't have harmonics) that the inverter puts back on the DC source.
A perfect DC source will have no voltage noise, but the current transients could be very high, depending on the filtering between the DC source and the inverter. As others have mentioned, the battery cable impedance is important.

 

A perfect DC source will have no voltage noise, but the current transients could be very high, depending on the filtering between the DC source and the inverter. As others have mentioned, the battery cable impedance is important.

Yes ,this is exactly what i'm after, 'DC noise'! It is the disturbance injected back into the pure DC source by inverter switching. I need to understand the waveforms of the same to gain an insight into the topic.
Thanx.

 

Tell us more about this inverter. Is it's output a so-called "modified sine wave", which is really a square wave with some dead time? Or, is it outputting a fairly good sine wave, using PWM modulation? Or, is it the "stepped approximation" type (such as the Xantrex SW4024)?

Are you only interested in the high frequency harmonics, due to rapid switching edges, with frequencies up in the radio frequency range? Or, are you also interested in the harmonics of the line frequency (perhaps up to the 51st harmonic) caused by the inductance of the battery cable?

 

sorry, i didn't mention earlier. it's not an electronics project, but an assumption made by me to understand the DC noise concept.
Thanx

 

sorry, i didn't mention earlier. it's not an electronics project, but an assumption made by me to understand the DC noise concept.
Thanx

In that case, you are just trying to understand the classic problem of supplying DC to a dynamically changing load. This is why we distribute filter caps all over circuits boards. All traces have resistance and inductance. A dynamically changing load current is directly converted to voltage changes at the end of the trace. A filter cap place at the device power terminals acts as a low pass filter to reduce the noise.

This effect can be studied with simple models of a DC source with source resistance and source inductance, with a filter cap.

 

This effect can be studied with simple models of a DC source with source resistance and source inductance, with a filter cap.

I understand, but i've this thing for pictorial representations! I grasp concepts better with diagrams and figures.
Thanx.

 

Someone plz help me! Thanx

 

Someone plz help me! Thanx

It's very difficult to give an answer of what the waveforms would look like off the top of our head. Also, you have not answered The Electrician's questions about the system details, which will surely impact the result.

We can talk about the general effect (as we have already), but quantifying and drawing plots is difficult without details. Even with details it is not trivial to give a good answer without doing some work. (although sometimes people here do quite a bit a work to help other people)

This is probably a case where you will have to help yourself mostly. There are a few ways to get at an answer. For example:

1. Make a numerical model of a typical inverter with a realistic DC driving source model. This means modeling a typicial PWM algorithm on transistor switches and various AC loads. Spice or Simulink are good tools to do this, although others are available.

2. Build a real inverter, or find/buy one and do measurements.

3. Do a research literature search for theory and/or measurements on this subject and hope someone has already worked out your answer.

I often run into this dilemma. I sometimes start with a literature search which is helpful. However, in the end I end up modeling the system to get the answers. There is quite a bit of literature about harmonics on the AC side, but even here, you would need to know the system approach to get the right answers. I'm not sure about the literature available about DC noise since I've not tried to research that before. Although, the basic cause of the DC noise is well-known, and proper care in cable design and filtering usually provides an adequate solution. In addition, inverters often use a DC link voltage sensor and use the instantaneous value of the voltage when computing the PWM commands. This type of feedback minimized the effect that DC variations have on the output, but does not remove the existing DC noise.

 

Surely any waveforms reflected back to a DC source will be damped to a lesser or greater degree depending upon the internal resistance of that source.

So you must start there.

 

Thanx very much for reply. Actually, the query is general in nature and not a project. But, I'll try to work out this problem on my own as per guidance given by you guys. Let's see how it turns out.

Best regards,
Shahvir